A generic tool for metadata manipulation of ecoacoustics audio recordings
Currently every environmental sensor captures audio recordings and recording metadata in a differant way.
There are efforts underway to standardize this process, but even in a perfect world, there are still plenty of problems to deal with:
- standards adoption takes time
- there a millions of recording made using older sensors
- there a many problems and quirks with existing sensors
EMU aims to be a babelfish—an adapter—between these formats. EMU can:
- extract metadata from audio recordings
- recognize and parse different date stamp formats
- rename files so that they have a consistent format
- fix problems in recordings so you can recover and use the data (idempotently)
- do this in various formats (human friendly, compact, json, json-lines, and csv)
- and is fast 🔥
For documentation please look in our documentation folder!
For example usage, please:
EMU has progressed to be a beta-level product.
EMU supports:
- Frontier Labs sensors (for FLAC and WAVE files)
- Wildlife Acoustics sensors (tested with SM3, SM4, & SM4BAT files)
- Open Acoustics devices
We want to support more files and formats. If you need one not covered here, then get in touch!
However, EMU is being actively used in large-scale automated ecoacoustics pipelines to validate and repair faulty audio recordings.
- EMU runs on Windows, Linux, and Mac*
- A docker container is provided (see Docker Hub
)
- The metadata extraction is quite good now
- There are several fixes implemented, but several more are in the backlog
*EMU needs special permissions to run on a Mac. See this article for guidance.
This project is funded through QUT Ecoacoustics and the Open Ecoacoustics projects
Special thanks also go to Doug Hynes and ECCC for sponsoring the development of the metadata extraction features.
